Recherche avancée

Sur d’autres sites (9797)

• arm : vp9itxfm : Skip empty slices in the first pass of idct_idct 16x16 and 32x32

  9 janvier 2017, par Martin Storsjö

  arm : vp9itxfm : Skip empty slices in the first pass of idct_idct 16x16 and 32x32
  This work is sponsored by, and copyright, Google.
  Previously all subpartitions except the eob=1 (DC) case ran with
  
  the same runtime :
                                       Cortex A7       A8       A9      A53
  
  vp9_inv_dct_dct_16x16_sub16_add_neon :   3188.1   2435.4   2499.0   1969.0
  
  vp9_inv_dct_dct_32x32_sub32_add_neon :  18531.7  16582.3  14207.6  12000.3
  By skipping individual 4x16 or 4x32 pixel slices in the first pass,
  
  we reduce the runtime of these functions like this :
  vp9_inv_dct_dct_16x16_sub1_add_neon :     274.6    189.5    211.7    235.8
  
  vp9_inv_dct_dct_16x16_sub2_add_neon :    2064.0   1534.8   1719.4   1248.7
  
  vp9_inv_dct_dct_16x16_sub4_add_neon :    2135.0   1477.2   1736.3   1249.5
  
  vp9_inv_dct_dct_16x16_sub8_add_neon :    2446.7   1828.7   1993.6   1494.7
  
  vp9_inv_dct_dct_16x16_sub12_add_neon :   2832.4   2118.3   2266.5   1735.1
  
  vp9_inv_dct_dct_16x16_sub16_add_neon :   3211.7   2475.3   2523.5   1983.1
  
  vp9_inv_dct_dct_32x32_sub1_add_neon :     756.2    456.7    862.0    553.9
  
  vp9_inv_dct_dct_32x32_sub2_add_neon :   10682.2   8190.4   8539.2   6762.5
  
  vp9_inv_dct_dct_32x32_sub4_add_neon :   10813.5   8014.9   8518.3   6762.8
  
  vp9_inv_dct_dct_32x32_sub8_add_neon :   11859.6   9313.0   9347.4   7514.5
  
  vp9_inv_dct_dct_32x32_sub12_add_neon :  12946.6  10752.4  10192.2   8280.2
  
  vp9_inv_dct_dct_32x32_sub16_add_neon :  14074.6  11946.5  11001.4   9008.6
  
  vp9_inv_dct_dct_32x32_sub20_add_neon :  15269.9  13662.7  11816.1   9762.6
  
  vp9_inv_dct_dct_32x32_sub24_add_neon :  16327.9  14940.1  12626.7  10516.0
  
  vp9_inv_dct_dct_32x32_sub28_add_neon :  17462.7  15776.1  13446.2  11264.7
  
  vp9_inv_dct_dct_32x32_sub32_add_neon :  18575.5  17157.0  14249.3  12015.1
  I.e. in general a very minor overhead for the full subpartition case due
  
  to the additional loads and cmps, but a significant speedup for the cases
  
  when we only need to process a small part of the actual input data.
  In common VP9 content in a few inspected clips, 70-90% of the non-dc-only
  
  16x16 and 32x32 IDCTs only have nonzero coefficients in the upper left
  
  8x8 or 16x16 subpartitions respectively.
  This is cherrypicked from libav commit
  
  9c8bc74c2b40537b0997f646c87c008042d788c2.
  Signed-off-by : Michael Niedermayer <michael@niedermayer.cc>
  

  • [DH] libavcodec/arm/vp9itxfm_neon.S
  • [DH] tests/checkasm/vp9dsp.c

• Checkasm : assembly testing and benchmarking tool

  11 juillet 2015, par Henrik Gramner

  Checkasm : assembly testing and benchmarking tool
  It provides the following features :
  
   * verify correctness by comparing output to the C version.
  
   * detect failure to save and restore clobbered callee-saved registers.
  
   * detect 32-bit parameters being used as if they were 64-bit in x86-64
  
     (the upper halves are not guaranteed to be zero - but in practice
  
     they very often are, which makes those bugs hard to spot otherwise).
  
   * easy benchmarking.
  Compile by running ’make checkasm’.
  
  Execute by running ’tests/checkasm/checkasm’.
  Optional arguments are ’—bench’ to run benchmarks for all functions,
  
  ’—bench=<pattern>’ to run benchmarks for all functions that starts with
  
  <pattern>, and ’<integer>’ to seed the PRNG for reproducible results.
  Contains unit tests for most h264pred functions to get started, more tests
  
  can be added afterwards using those as a reference.
  Loosely based on code from x264. Currently only supports x86 and x86-64,
  
  but additional architectures shouldn’t be too much of an obstacle to add.
  Note that functions with floating point parameters or floating point
  
  return values are not supported. Some compiler-specific features or
  
  preprocessor hacks would likely be required to add support for that.
  Signed-off-by : Janne Grunau <janne-libav@jannau.net>
  

  • [DBH] .gitignore
  • [DBH] tests/Makefile
  • [DBH] tests/checkasm/Makefile
  • [DBH] tests/checkasm/checkasm.c
  • [DBH] tests/checkasm/checkasm.h
  • [DBH] tests/checkasm/h264pred.c
  • [DBH] tests/checkasm/x86/Makefile
  • [DBH] tests/checkasm/x86/checkasm.asm

• dnn_backend_native_layer_mathunary : add floor support

  6 août 2020, par Mingyu Yin

  dnn_backend_native_layer_mathunary : add floor support
  It can be tested with the model generated with below python script :
  import tensorflow as tf
  
  import os
  
  import numpy as np
  
  import imageio
  
  from tensorflow.python.framework import graph_util
  
  name = 'floor'
  pb_file_path = os.getcwd()
  
  if not os.path.exists(pb_file_path+'/{}_savemodel/'.format(name)) :
  
      os.mkdir(pb_file_path+'/{}_savemodel/'.format(name))
  with tf.Session(graph=tf.Graph()) as sess :
  
      in_img = imageio.imread('detection.jpg')
  
      in_img = in_img.astype(np.float32)
  
      in_data = in_img[np.newaxis, :]
  
      input_x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
  
      y_ = tf.math.floor(input_x*255)/255
  
      y = tf.identity(y_, name='dnn_out')
  
      sess.run(tf.global_variables_initializer())
  
      constant_graph = graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
      with tf.gfile.FastGFile(pb_file_path+'/{}_savemodel/model.pb'.format(name), mode='wb') as f :
  
          f.write(constant_graph.SerializeToString())
      print("model.pb generated, please in ffmpeg path use\n \n \
  
      python tools/python/convert.py {}_savemodel/model.pb —outdir={}_savemodel/ \n \nto generate model.model\n".format(name,name))
      output = sess.run(y, feed_dict= input_x : in_data)
  
      imageio.imsave("out.jpg", np.squeeze(output))
      print("To verify, please ffmpeg path use\n \n \
  
      ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.pb:input=dnn_in:output=dnn_out:dnn_backend=tensorflow -f framemd5 {}_savemodel/tensorflow_out.md5\n  \
  
      or\n \
  
      ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.pb:input=dnn_in:output=dnn_out:dnn_backend=tensorflow {}_savemodel/out_tensorflow.jpg\n \nto generate output result of tensorflow model\n".format(name, name, name, name))
      print("To verify, please ffmpeg path use\n \n \
  
      ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.model:input=dnn_in:output=dnn_out:dnn_backend=native -f framemd5 {}_savemodel/native_out.md5\n  \
  
      or \n \
  
      ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.model:input=dnn_in:output=dnn_out:dnn_backend=native {}_savemodel/out_native.jpg\n \nto generate output result of native model\n".format(name, name, name, name))
  Signed-off-by : Mingyu Yin <mingyu.yin@intel.com>
  

  • [DH] libavfilter/dnn/dnn_backend_native_layer_mathunary.c
  • [DH] libavfilter/dnn/dnn_backend_native_layer_mathunary.h
  • [DH] tests/dnn/dnn-layer-mathunary-test.c
  • [DH] tools/python/convert_from_tensorflow.py
  • [DH] tools/python/convert_header.py